Global phase diagrams for freezing in porous media

Abstract

Effect of the fluid-wall interaction on freezing of confined fluids: Toward the development of a global phase diagramUsing molecular simulations and free energy calculations based on Landau theory, we show that freezing/melting behavior of fluids of small molecules in pores of simple geometry can be understood in terms of two main parameters: the pore width H* ͑expressed as a multiple of the diameter of the fluid molecule͒ and a parameter ␣ that measures the ratio of the fluid-wall to the fluid-fluid attractive interaction. The value of the ␣ parameter determines the qualitative nature of the freezing behavior, for example, the direction of change in the freezing temperature and the presence or absence of new phases. For slit-shaped pores, larger ␣ values lead to an increase in the freezing temperature of the confined fluid, and to the presence of a hexatic phase. For pores that accommodate three or more layers of adsorbate molecules several kinds of contact layer phase ͑inhomogeneous phases in which the contact layer has a different structure than the inner layers͒ are observed. Smaller ␣ values lead to a decrease in the freezing temperature. The parameter H* determines the magnitude of shift in the freezing temperature, and can also affect the presence of some of the new phases. Results are presented as plots of transition temperature vs ␣ for a particular pore width. Experimental results are also presented for a variety of adsorbates in activated carbon fibers ͑ACF͒ covering a wide range of ␣ values; the ACF have slit-shaped pores with average pore width 1.2 nm. The experimental and simulation results show qualitative agreement.

Effect of Pressure on the Freezing of Pure Fluids and Mixtures Confined in Nanopores

“…Upon reducing the width of the confined space to approach the range of the intermolecular forces, significant shifts in the freezing temperature are observed, and in some cases, new surface-or confinement-induced phases occur. 2,11,12 Previous experimental, molecular simulation, and theoretical studies have shown that, for simple fluids and pore geometries, the freezing temperature can be described as a function of the reduced pore size H* ) H/σ (H is the pore width, and σ the diameter of an adsorbate molecule) and the ratio of the wall/fluid (wf) to the fluid/fluid (ff) interactions, R ∼ CF w ε wf / ε ff , where F w and ε are the density of wall atoms and the potential well depth, respectively, and C is a constant that depends on the wall geometry. The freezing temperature T f is decreased compared to the bulk value, T f 0 , for R < ∼1, while it is increased for R > ∼1.…”

The self-referential method combined with thermodynamic integration

“…In terms of efficiency, it should be similar to other methods based on thermodynamic integration, such as the lattice coupling expansion method of Frenkel and Ladd 20 and Meijer et al 27 This work has shown that it is much more efficient than the previous method based on parameter hopping, and we also expect it to be much more efficient than other biased sampling methods. 2,9,12,22,37 In this work each single-sized system consists of a cubic simulation box, so the doublesized system is a rectangular box. However, because all primary crystal unit cells are parallelepipeds, the SR method should be completely general.…”

“…It is in principle exact ͑to within statistical error͒ provided suitable choices of ␣ 1 , ␣ m , n ␣ , and n b are made, it should always avoid problems associated with phase transitions, 18,26 and it avoids using the grand-canonical ensemble. [8][9][10][11][12][13][14][15] In terms of convenience, there is no need to evaluate a complicated "center-of-mass" correction or the free energy of a reference crystal, 3,19,20,27,35,36 or search for optimal parameters for reference states, there is no need to modify the method for hard-core molecules, 3 no need to devise optimal paths on a case-by-case basis along integration parameters that avoid phase transitions, 23 and we expect there is no need to "integrate away" pore walls, 16 in the case of confined crystals, for example, so that an integration path that connects with an ideal reference crystal can be defined ͑although further work is needed to confirm this͒. Any molecular simulation method that allows configurational Hamiltonians of the form ͑5͒ can be used.…”

“…Despite all this, there are several examples [8][9][10][11][12][13][14][15] in the literature where the grand-canonical ensemble has been used to simulate space-filling crystalline solids and 2D crystals on the inner surfaces of slit pores without box length fluctuations. All this work should be considered carefully because in every case these are not equilibrium simulations.…”

Wetting of Nanostructurized Sapphire and Gold Surfaces

“…It was already shown that the pore size and geometry have significant influence on the phenomena occurring inside pores such as phase transitions and chemical reactions [5][6][7] and that contact angle can be correlated with the microscopic wetting parameter α [8].…”